Exploring CeO₂-Doped Co/SBA-15 Catalysts for Acetic Acid Oxidative Steam Reforming

Abstract

This work explores the effect of the incorporation of CeO₂ into Co/SBA-15 catalysts in hydrogen production through acetic acid oxidative steam reforming as a bio-oil aqueous phase model compound. CeO₂ was incorporated (5–30 wt.%) to improve the physicochemical properties of the catalyst. XRD analysis confirmed that the addition of CeO₂ resulted in smaller Co⁰ mean crystallite sizes, while H₂-TPR showed enhanced reducibility properties. The catalytic performance was evaluated in the 400–700 °C range, S/C molar ratio = 2, O₂/C molar ratio = 0.0375, WHSV = 30.2 h⁻¹, and P = 1 atm. Catalysts containing 10 and 20 wt.% of CeO₂ exhibited the best catalytic performance, achieving nearly complete conversions and H₂ yield values, approaching thermodynamic equilibrium at 550 °C. Both samples maintained an acetic acid conversion above 90% after 30 h of time-on-stream, with H₂ yields above 55% along the steam reforming tests. This agrees with their lower coke formation rates (7.2 and 12.0 mg_coke·g_cat⁻¹·h⁻¹ for Co/10CeO₂-SBA15 and Co/20CeO₂-SBA15, respectively).

1. Introduction

Traditional fuels have been the starting point of industrial and energy development worldwide. However, burning fossil fuels significantly contributes to the release of greenhouse gases, driving climate change. CO₂ emissions and other pollutants increase global temperatures, leading to extreme climate, melting glaciers, and rising sea levels [1,2]. Given this alarming reality, it is imperative to implement robust environmental policies to lessen emissions and promote the transition in the direction of more sustainable sources of energy. To address these challenges, the established aim is to limit the global temperature increase to 1.5 °C until the end of the century [3,4,5]. Consequently, governments have implemented environmental policies that promote the utilization of renewable energies, the development of energy efficiency, and the integration of clean technologies [6]. In these circumstances, hydrogen has risen as a promising alternative, considering its high energetic potential and environmental advantages [7]. Hydrogen production from renewable resources, such as biomass, will be a sustainable alternative for managing organic waste and reducing the dependency on fossil fuels, playing a potentially key role in the industrial and transport sectors’ decarbonization [7,8].

Thermochemical processes, such as pyrolysis, use biomass to generate valuable products such as syngas, bio-char, and bio-oil. The pyrolysis bio-oil could be composed of organic as well as aqueous phases [8,9,10]. The organic phase consists of high-energy compounds, which could be further enhanced through hydrotreating to produce biofuels. However, the aqueous phase has a lower energy potential than the organic phase, given that it consists mainly of water and several oxygenated compounds with low and high molecular weight [8]. Despite having a lower energy value, the bio-oil aqueous phase could be revalorized via catalytic steam reforming, producing hydrogen [9,10,11,12,13]. Given the heterogeneity of the bio-oil aqueous phase composition, many authors focus their investigation on representative compounds, such as acetic acid, since it represents approximately 20 wt.% of these organic compounds [14,15,16,17].

In acetic acid steam reforming, steam reacts with this organic compound at high temperatures to produce hydrogen, as represented in Equation (1). This process is highly endothermic, requiring high energy consumption to maintain the high temperatures required.

$${\mathrm{C}}_2{\mathrm{H}}_4{\mathrm{O}}_2{ + 2\mathrm{H}}_2\mathrm{O}\leftrightarrow {4 \mathrm{H}}_2{ + 2\mathrm{CO}}_2$$

(1)

In this scenario, some authors have reported that the inclusion of oxygen in the feed stream leads to oxidative steam reforming [18,19]. The primary advantage of this configuration is that exothermic oxidation reactions provide part of the energy requirements to drive forward the endothermic steam reforming reactions. An oxidizing environment would also reduce the coke deposited on the catalyst by its combustion. However, hydrogen generation will be slightly reduced, as shown in Equation (2).

$${\mathrm{C}}_2{\mathrm{H}}_4{\mathrm{O}}_2+\left({\displaystyle \frac{\mathrm{n}}{2}}\right){\mathrm{O}}_2+\left(2-\mathrm{n}\right){ \mathrm{H}}_2\mathrm{O}\leftrightarrow {(4-\mathrm{n}) \mathrm{H}}_2{+2\mathrm{CO}}_2$$

(2)

Apart from the main reaction, secondary reactions can occur, such as a water–gas shift (Equation (3)), the thermal decomposition of acetic acid (Equation (4)), ketonization (Equation (5)), acetone steam reforming (Equation (6)), methane steam reforming (Equation (7)), coke formation (Equations (8) and (9)), and the Boudouard reaction (Equation (10)) [20].

$${\mathrm{CO}+\mathrm{H}}_2\mathrm{O} \leftrightarrow { \mathrm{H}}_2{+\mathrm{CO}}_2$$

(3)

$${\mathrm{C}}_2{\mathrm{H}}_4{\mathrm{O}}_2 \leftrightarrow { \mathrm{CH}}_4{+\mathrm{CO}}_2$$

(4)

$${2 \mathrm{C}}_2{\mathrm{H}}_4{\mathrm{O}}_2 \leftrightarrow { \mathrm{C}}_3{\mathrm{H}}_6{\mathrm{O}+\mathrm{CO}}_2{+\mathrm{H}}_2\mathrm{O}$$

(5)

$${\mathrm{C}}_3{\mathrm{H}}_6{\mathrm{O}+5\mathrm{H}}_2\mathrm{O}\leftrightarrow {8 \mathrm{H}}_2{+3\mathrm{CO}}_2$$

(6)

$${\mathrm{CH}}_4{+2\mathrm{H}}_2\mathrm{O} \leftrightarrow { 4 \mathrm{H}}_2{+\mathrm{CO}}_2$$

(7)

$${\mathrm{CO}}_2{+2 \mathrm{H}}_2 \leftrightarrow {2 \mathrm{H}}_2\mathrm{O}+\mathrm{C}$$

(8)

$${\mathrm{CO}+\mathrm{H}}_2 \leftrightarrow {\mathrm{H}}_2\mathrm{O}+\mathrm{C}$$

(9)

$$2 \mathrm{CO} \leftrightarrow {\mathrm{CO}}_2+\mathrm{C}$$

(10)

The development of a suitable catalyst to promote selectivity to hydrogen formation is essential because it influences the reaction mechanism and thus is able to enhance the reaction rate toward hydrogen. Catalyst deactivation is still one of the principal issues in the scientific community when referring to reforming catalysts [21] due to the carbon formation or active phase sintering, and in the context of oxidative steam reforming, where an oxidizing environment is present, the active phase reoxidation [18].

Regarding the catalyst formulation, transition metals such as cobalt have been widely used as the active phase [15,22,23,24] in reforming reactions against noble metals, given their promising activity in breaking down C-C and C-H bonds and lower prices. Furthermore, the election of adequate support is crucial for enhancing the active phase dispersion. Previous works [16,21,25] have demonstrated that materials based on mesoporous SiO₂, such as SBA-15, used as support, allow higher active phase dispersion due to their highly ordered mesoporous structure with a large surface area. Additionally, it has been reported that the incorporation of CeO₂ into the catalyst formulation improved its activity and resistance toward coke deposition, by the capability of storing, releasing, and transporting oxygen [26].

CeO₂-SiO₂ supports have been previously studied in steam reforming reactions. In this respect, Calles et al. [27] researched CeO₂-doped Ni/SiO₂ (Ni/CS) and Ni/SiO₂ (Ni/S) catalysts synthesized via incipient wet impregnation. In this study, it was demonstrated that the H₂-TPR profiles indicated a slightly lower temperature peak corresponding to NiO reduction for the Ni/CS catalyst, due to the oxygen vacancies in CeO₂, thus facilitating higher reducibility. In catalytic tests on ethanol steam reforming, it was demonstrated that a CeO₂-doped Ni/SiO₂ catalyst enhanced the catalytic activity, maintaining complete ethanol conversion over time. Cifuentes et al. [28] studied Rh-Pt-based catalysts supported on CeO₂-SiO₂ for ethanol steam reforming. The CeO₂-SiO₂ supports were synthesized by the co-precipitation method, and the active phase was incorporated by the incipient wetness co-impregnation method. They determined that the optimum Si content was 33 mol%, equivalent to a Si:Ce ratio of 2, whereby the highest catalytic yield was obtained, reaching 5.3 mol H₂/mol EtOH at 700 °C. Ribeiro et al. [29] evaluated Ce_0.65Me_0.1Si_0.25 catalysts (Me = V, Cr, Mn, Nb, W, Mo) in the production of oxygenates from ethanol/steam mixtures. They reported that in CeO₂-SiO₂ supports, the interactions between silica and surface ceria nanocrystals lead to Ce-O-Si unit formation, promoting increased surface area. Additionally, in Mn and Cr-doped catalysts, an improvement in reducibility was observed, attributed mainly to an increase in CeO₂ oxygen mobility, thus enhancing the catalytic activity. Results showed an initial ethanol conversion close to 90%, with product distribution dominated by hydrogen and a suppressed ethylene formation, promoting a higher catalytic stability for 6 h.

Palma et al. [30] investigated the ethanol oxidative steam reforming reaction using Ni and Pt-Ni catalysts supported on CeO₂-SiO₂, achieving complete ethanol conversions and hydrogen yields close to thermodynamic equilibrium at temperatures above 450 °C. These works underscore the potential of CeO₂-SiO₂ supports, highlighting that the incorporation of CeO₂ into the SiO₂ matrix enhances oxygen mobility and strengthens metal–support interactions. This promotes the dispersion of the active phase and, likewise, improves the catalytic stability.

The current employment of mesostructured SiO₂, such as SBA-15, enhances these structural and catalytic advantages. Several studies had explored CeO₂-doped SBA-15 in the steam reforming of ethanol [31], methanol [32], propylene glycol [33], glycerol [34], and toluene [35], thus providing evidence of its viability for hydrogen production from oxygenated compounds. This raises potential applications in the upgrading of the aqueous phase of bio-oil, focusing on acetic acid conversion by reforming reactions.

Acetic acid oxidative steam reforming (AAOSR) has been explored mainly at autothermal conditions (ATR). Research such as that of Su et al. [36] evaluated the effect of Y-doping Ni/CeO₂ catalysts in ATR at 700 °C, with a molar ratio of HAc/H₂O/O₂/N₂ = 1/4/0.28/3, and GHSV = 50,600 mL/(g·h). Results showed that the NCY10 catalyst (14.4 wt% NiO, 75.7 wt% CeO₂, 9.9 wt% YO_1.5) achieved complete conversions during 10 h of reaction, with a 2.6 mol-H₂/mol AcOH stable yield. In addition, for the catalyst used, TGA analysis did not show evidence of coke formation, suggesting a high stability of the system. Furthermore, Zhou et al. [37] synthesized mesoporous Ni-xSm-Al-O catalysts to evaluate the Sm doping effect on the catalyst. A high surface area, strong basicity and weak acidity, and highly dispersed Ni⁰ crystals with a mean crystallite size of 9.9 nm were observed after doping with 2% Sm. These characteristics enhanced the catalytic activity (T = 700 °C, P = 1 atm; GHSV = 30,000 mL/(g·h), and a molar ratio of HAc/H₂O/O₂/N₂ = 1/4/0.28/3.90), with 100% acetic acid and H₂ yield conversions of 2.6 mol-H₂/mol AcOH achieved during 30 h of reaction. Some recent work by Megía et al. [25,38] was focused on nickel-based catalysts with different supports, highlighting the Ni/Ce-m catalyst, which achieved conversions around 97% during 5 h of reaction (T = 500 °C, P = 1 atm; GHSV = 11,000 h⁻¹, and a molar ratio HAc/H₂O/O₂ = 1/2/0.08). Nevertheless, a coke formation of 120.6 mg_coke·g_cat⁻¹·h⁻¹ was reported, thus leading to the La₂O₃ addition, enhancing the dispersion of the metallic phase as well as significantly lowering the carbon deposition to 62.3 mg_coke·g_cat⁻¹·h⁻¹. Lastly, Co-Rh/CeO₂ bimetallic catalysts [39] exhibited long-term stability during 25 h (T = 600 °C, P = 1 atm; WHSV = 30.1 h⁻¹, and a molar ratio HAc/H₂O/O₂ = 1/2/0.08). The studies highlight the relevance of AAOSR and the importance of further optimization of the catalyst formulation to improve process efficiency, selectivity toward H₂, and stability against deactivation.

Although CeO₂ has been widely recognized for enhancing catalytic dispersion and coke resistance, most previous studies have focused on Ni-based catalysts on reforming reactions under autothermal conditions. Based on the above, the effect of the incorporation of CeO₂ in the catalytic performance of a Co-based catalyst supported over SBA-15 in the AAOSR was evaluated. An optimal CeO₂ proportion would enhance the cobalt reducibility, dispersion, and metal–support interaction, enhancing an efficient catalytic performance for hydrogen production.

2. Materials and Methods

2.1. Catalyst Synthesis

Commercial mesoporous silica (SBA-15) purchased from ACS Material was used as a support since its high surface area and hexagonal ordered structure allows better active phase dispersion than amorphous supports. CeO₂-modified SBA-15 supports were synthesized following the incipient wetness impregnation technique using aqueous solutions of Ce(NO₃)₃·6H₂O (99%, Sigma-Aldrich, St. Louis, MO, USA), as the Ce precursor, in different concentrations to obtain different Ce contents in the modified support: 5, 10, 20, and 30 wt.%. After that, materials were calcined in airflow at 550 °C for 5.5 h using a 1.8 °C/min heating rate.

The active phase was incorporated according to the same procedure explained above, using Co(NO₃)₂·6H₂O (>98%, Sigma-Aldrich, St. Louis, MO, USA) solutions to get 7 wt.% of Co in the synthesized catalysts. The synthesized catalysts have been denoted as Co/xCeO₂-SBA-15, where x indicates the Ce theoretical loading (wt.%).

2.2. Catalyst Characterization

All the samples were characterized by multiple procedures. ICP-OES analysis was performed to measure the Ce and Co contents in the synthesized catalysts, employing an Agilent ICP-OES 5800 VDV spectrophotometer (Agilent, Santa Clara, CA, USA). Before analysis, the catalysts were dissolved by acid digestion (H₂SO_4—HF mixture). The textural properties were evaluated based on N₂ physisorption at 77 K using a Micromeritics TRISTAR 3000 sorptometer (Micromeritics, Norcross, GA, USA). Specific surface areas and pore sizes were estimated using the BET method and the BJH formula, respectively. X-ray diffractograms were acquired on a Philips X’PERT PRO diffractometer (Philips, Eindhoven, The Netherlands) equipped with Cu Kα radiation. Mean crystallite sizes were calculated using the Scherrer equation.

$$\mathrm{d}={\displaystyle \frac{\mathrm{K}·\lambda }{\beta ·\cos \theta }}$$

(11)

where “d (nm)” refers to the mean crystallite diameter, “K” is the Scherrer constant (0.9), “λ” is the X-ray wavelength (Cu-Kα at 0.15406 nm), “β” refers to the FWHM (full width at half maximum) of the highest intensity diffraction peak, and “θ” indicates the diffraction half-angle.

To assess the catalyst’s reducibility, H₂-TPR analyses were performed on a Micromeritics in Situ Catalysts Characterization System (ICCS, Micromeritics, Norcross, GA, USA), where a reducing gas mixture flowed through the sample (10% H₂/Ar, 60 mL/min) while heating at 10 °C/min up to 900 °C.

The morphology of the prepared materials was evaluated using TEM in a 120 kV JEOL JEM-1400 FLASH microscope (JEOL, Tokyo, Japan); the samples were previously dispersed in acetone. Using a Cary series UV-Vis-NIR spectrophotometer (Agilent, Santa Clara, CA, USA) at wavelength range of 200–800 nm, the CeO₂ oxidation stages on the synthesized materials were evaluated. Finally, TGA and Raman spectra were used to assess the coke deposition on the used catalysts. For that purpose, a TGA-DSC Mettler Toledo thermobalance (Mettler Toledo, Greifensee, Switzerland) employing airflow at a 5 °C/min heating rate until 1000 °C and a Horiba LabRam HR evolution spectrometer (HORIBA Scientific, Kyoto, Japan) with a 532 nm laser beam (nominal power 200 mW) were utilized.

2.3. Catalytic Performance Tests

The AAOSR reactions (P= 1 atm, T = 400–700 °C) were carried out on a MICROACTIVITY-PRO unit (PID Eng&Tech. S.L., Alcobendas, Madrid, Spain), whose schematic diagram can be found elsewhere [25]. Co/xCeO₂-SBA-15 catalysts (300 mg) were loaded inside the reactor and with an in situ reduction by pure hydrogen flow (30 mL/min) for 30 min at 600 °C before the reaction. After the catalyst activation, a pure nitrogen stream was fed into the reactor until the reaction conditions were reached. A Gilson 307 piston pump (Gilson, Middleton, WI, USA) was used to feed the acetic acid aqueous solution to the reactor (steam-to-carbon molar ratio = 2) at a WHSV of 30.2 h⁻¹ (WHSV was defined as the mass flow rate of reactants—acetic acid, water, oxygen, and nitrogen—divided by the mass of the loaded catalyst in the reactor). The O₂ supply was set to operate with an O₂/C molar ratio of 0.0375, maintaining the nitrogen/oxygen mixture’s total flow at 60 mL/min. Condensable reaction vapors were condensed at 4 °C, collected, and subsequently analyzed on an Agilent 7820A GC System (Agilent, Santa Clara, CA, USA) while the non-condensable gaseous stream was analyzed in line on an Agilent MicroGC 490 (Agilent, Santa Clara, CA, USA). The mean value of three measurements of each gas sample was used. Figure 1 shows a simplified schematic of the Microactivity—Pro unit, focusing on the feed section, reaction section, and product analyses.

The reaction performance was evaluated through acetic acid conversion (X), H₂ yield (Y_H2), and carbon-containing product selectivities (S_CO, SCH₄, SCO₂, SC₃H₆O), calculated as shown in Equations (11)–(13):

$${\mathrm{X}}_{\mathrm{reactant}}\left(\%\right)={\displaystyle \frac{{\mathrm{F}}_{\mathrm{reactant},\mathrm{in}}{-\mathrm{F}}_{\mathrm{reactant},\mathrm{out}}}{{\mathrm{F}}_{\mathrm{reactant},\mathrm{in}}}}·100$$

(12)

$${\mathrm{Y}}_{{\mathrm{H}}_2}\left(\%\right)={\displaystyle \frac{{\mathrm{F}}_{{\mathrm{H}}_2,\mathrm{out}}}{{\mathrm{n}}_{\mathrm{i}}{.\mathrm{F}}_{\mathrm{reactant},\mathrm{in}}}}·100$$

(13)

$${\mathrm{S}}_{\mathrm{i},\mathrm{carbon}-\mathrm{containing} \mathrm{products}}\left(\%\right)={\displaystyle \frac{{\mathrm{F}}_{\mathrm{i},\mathrm{carbon}-\mathrm{containing} \mathrm{products}}}{{\mathrm{a}}_{\mathrm{i}}.\left({\mathrm{F}}_{\mathrm{i},\mathrm{in}}{-\mathrm{F}}_{\mathrm{i},\mathrm{out}}\right)}}·100$$

(14)

where “F” refers to the molar flow rates of “i” species at either the inlet (in) or the outlet (out) of the reactor, “n_i” represents the stoichiometric coefficient for maximum hydrogen generation, and “a_i” is the stoichiometric factor for carbon between the acetic acid and each compound.

3. Results and Discussion

3.1. Catalyst Characterization

The elemental composition and the textural characteristics of the catalysts synthesized in the current research are summarized in

Table 1. Physical–chemical characteristics of Co/xCeO₂-SBA-15 catalysts.

Sample	Ce a	Co a	SBET	Vp b	Dp c	DCo3O4 d	DCo0 e
	(wt.%)	(wt.%)	(m2/g)	(cm3/g)	(nm)	(nm)	(nm)
Co/SBA-15	-	7.1	490	0.90	8.8	10.1	6.1
Co/5CeO2-SBA-15	5.4	7.4	388	0.85	8.8	7.4	4.5
Co/10CeO2-SBA-15	12.3	7.4	344	0.77	8.9	6.8	4.2
Co/20CeO2-SBA-15	23.3	7.1	303	0.62	8.2	7.5	n.d. f
Co/30CeO2-SBA-15	33.0	7.0	241	0.50	8.2	7.6	n.d. f

^a Measured by ICP-OES in calcined catalysts. ^b Calculated at P/P₀ = 0.95. ^c BJH pore size distribution maximum. ^d Determined by the (311) reflecting plane of the Co₃O₄ pattern in XRD of calcined samples. ^e Determined by the (111) reflecting plane of the Co⁰ pattern in XRD of reduced samples. ^f Not detectable. Equipment detection range (>3 nm).

. Concerning metal loadings calculated via ICP-OES, the calcined catalysts have Ce and Co contents nearing the nominal value established at the SBA-15 and Ce/SBA-15 impregnation, respectively.

Figure 2A and Figure S1 show the N₂ physisorption isotherms of the calcined catalysts and supports, respectively. These isotherms exhibited a type IV isotherm curve with an H1-type hysteresis loop according to the IUPAC classification, which is typical for mesoporous material consisting of well-defined cylindrical pore channels [40]. These results demonstrate that the characteristic porous structure of bare SBA-15 was preserved after the support modification with ceria and the cobalt active phase incorporation. Notably, some differences in the isotherm shape can be appreciated with the increase in Ce content, indicating a slight modification of the mesoporous structure. This effect is particularly evident for Co/20CeO₂-SBA-15 and Co/30CeO₂-SBA-15 samples because of their higher Ce content, resulting in 38.2% and 50.8% decreases in BET surface area, respectively, compared to the Co/SBA-15 catalyst. In line with this assumption, the pore volume, BET surface area, and mean pore diameter displayed in Table 1 and Table S1, corresponding to catalyst and supports, respectively, became smaller as the CeO₂ amount increased, which can be ascribed to partial filling of SBA-15 channels. Similar behavior was observed after cobalt incorporation into the ceria-modified silica supports.

The UV-Vis diffuse reflectance spectra of the supports are shown in Figure S2. The UV-Vis spectra of the CeO₂-SBA-15 supports show a broad absorption band centered at about 288 nm, ascribed to charge transfer transitions from O²⁻ to Ce⁴⁺ [41,42]. An additional shoulder around 255 nm is noted in particular for the 5CeO₂-SBA-15 support, possibly indicating the presence of Ce³⁺ species.

Band gap energy was estimated by the Tauc method, calculated from the intercept with the abscissa in a linear regression of the (F(R)·hυ)^1/n vs. hυ representation, considering a direct electronic transition (n = ½) [43] (Figure S3). The obtained values exhibit a slight variation, 3.7 eV (5CeO₂-SBA-15), 3.5 eV (10CeO₂-SBA-15), 3.6 eV (20CeO₂-SBA-15), and 3.7 eV (30CeO₂-SBA-15). Such differences indicate possible alterations in the electronic structure of the supports, linked to the presence of structural defects. These defects are probably caused by the formation of oxygen vacancies. In particular, the decrease in the band gap observed in the 10CeO₂-SBA-15 and 20CeO₂-SBA-15 supports is likely related to an improved density of oxygen vacancies [44].

The crystalline phases of ceria and cobalt oxides supported on SBA-15 material were analyzed by XRD, and their diffraction patterns are illustrated in Figure 2B. All the catalysts show a broad peak between 20 and 30°, ascribed to the amorphous SiO₂ on the SBA-15 framework [45]. The intensity of the peak becomes smaller as the CeO₂ content increases. Peaks related to the Co₃O₄ cubic phase (JCPDS 01-071-4921) appear in all the patterns of the materials after calcination, showing reflections at 2θ = 31.2, 36.8, 44.8, 59.5, 65.4, and 77.7°, referring to the (220), (311), (400), (511), (440), and (533) diffraction planes, respectively. In addition to Co₃O₄ diffraction lines, Co/xCeO₂-SBA-15 samples also show reflections at 2θ = 28.5, 32.9, 47.5, 56.4, 66.5, and 79.1° according to the planes (111), (200), (220), (311), (400), and (420) of the cubic CeO₂ structure (JCPDS 01-089-8436), respectively. Comparing these XRD patterns, it is possible to observe that, as the CeO₂ content increases, Co₃O₄ peaks become broader and less intense, suggesting more dispersed Co-species over the xCeO₂-SBA-15-modified supports [46]. This effect was evidenced by the mean crystallite sizes on the Co₃O₄ main diffraction line. Over xCeO₂-SBA-15, smaller crystallite sizes were achieved (see Table 1) than over Co/SBA-15, thus indicating higher metal dispersion in the modified CeO₂ samples. Similarly, Jiménez-Morales et al. developed a variety of Ni/SBA-15 catalysts for the hydrogenolysis of glycerol, while studying the impact of Ce incorporation, and concluded that high active phase dispersion was achieved as a consequence of the interactions between metallic oxide particles and the cerium oxide ones [47].

The reducibility was assessed for catalysts produced using H₂-TPR analysis, the reducing profiles of which are depicted in Figure 3. In all cases, several reduction features can be distinguished. The Co/SBA-15 catalyst displayed three reduction stages at 344 °C, 393 °C, and 472 °C. The first two are associated with the reduction of Co³⁺ to Co²⁺ and afterwards to Co⁰, while the third one, centered at 472 °C, is related to the existence of Co oxide species with a strong engagement with the SBA-15 support [34]. Cerium-modified catalysts showed a low-temperature reduction peak between 347 °C and 361 °C, which could be described as a simultaneous Co³⁺ reduction to Co²⁺ and Co²⁺ reduction to Co⁰ [48]. This reduction performance could be correlated with the redox properties of CeO₂, which has a well-known oxygen storage and mobilization capacity [26]; this characteristic allowed it to capture and release oxygen easily, generating vacancies in the structure. Incorporating CeO₂ into the SBA-15 support facilitates the surface migration and oxygen transfer from cobalt oxides, thus favoring the reduction processes. Furthermore, in these samples, the peak centered at 472 °C shifts toward slightly higher temperatures, suggesting a stronger metal–support interaction (MSI) between the cobalt oxide species and the support. It is associated with the smaller Co crystallite formation, increasing the interfacial contact area, leading to the surface chemical bonding [49]. This highlights the significant role of metal–support interaction in catalyst design since a suitable support will interact effectively with the metallic species, minimizing its surface migration and preventing sinterization. In contrast, scarce metal–support interaction would promote particle agglomeration, leading to a negative impact on the catalytic performance [50]. The Co/10CeO₂-SBA-15 catalyst exhibited the most strongly pronounced reduction feature associated with strong metal–support interactions, while this peak was almost absent in the Co/30CeO₂-SBA-15 catalyst. It could be attributed to the slight modification of the mesoporous structure. A further decrease in surface area could probably promote CeO₂ agglomeration, hindering the effective cobalt oxide species dispersion. Consequently, the metal–support interaction effect between the cobalt phase and the SBA-15 structure was reduced. Additionally, reduction features at higher temperatures (735–769 °C) arise in all Co/xCeO₂-SBA-15 catalysts, ascribed to the partial reduction of CeO₂ to CeO_2−x [25,51]. The reduction degree of Co species was assessed by quantifying the H₂ consumption throughout the TPR analysis by comparing the experimental and theoretical H₂ uptake (

Table 2. Experimental and theoretic H₂ consumption and Co reduction degree for Co/xCeO₂-SBA-15 calcined catalysts.

Sample	Theoretical Uptake	Experimental Uptake	Degree of Co Reduction
	(µmol H2/gcat)	(µmol H2/gcat)	(%)
Co/SBA-15	1606	1530	95
Co/5CeO2-SBA-15	1674	1642	98
Co/10CeO2-SBA-15	1683	1679	~100
Co/20CeO2-SBA-15	1604	1554	97
Co/30CeO2-SBA-15	1572	1508	96

). The theoretical hydrogen uptake was calculated from the full reduction stoichiometry of the cobalt oxide loading (measured by ICP-OES). Meanwhile, the H₂ experimental uptake was obtained by the H₂-TPR profiles’ integration of the hydrogen uptake peaks, corresponding to the Co₃O₄ reduction. All the catalysts exhibited a reduction degree above 95%, with slightly higher values observed for the Ce-doped samples, which is in agreement with the shift in the reduction profile toward lower temperatures for Co species. This slight increase in the reducibility can be attributed to the oxygen vacancy generation caused by the incorporation of ceria [26]. In this respect, the Co/10CeO₂-SBA-15 sample reached the highest reduction level, with its experimental uptake, in terms of μmol H₂·g_cat⁻¹, closely matching the theoretical value.

After the catalysts’ reduction step at 600 °C, samples were characterized by XRD analysis, and the resultant diffraction patterns are displayed in Figure 4. As can be discerned, no diffraction peaks corresponding to Co oxides can be distinguished, confirming that the reduction step proceeded successfully. In this sense, peaks arising at 2θ = 44.1 and 76.4° are attributed to the cubic phase of Co⁰ (JCPDS-15-0806). The Co⁰ mean crystal sizes in the reduced catalysts were determined based on the principal Co⁰ diffraction plane (111) using the Scherrer equation (Table 1). Notably, no diffraction features corresponding to Co⁰ can be observed for Co/20CeO₂-SBA-15 and Co/30CeO₂-SBA-15 catalysts, which reached a mean crystallite size lower than the detectable limit of the technique. This suggests the metallic phase was widely dispersed on the SBA-15 support modified with 20–30 wt.% of ceria.

TEM micrographs for the reduced catalysts are presented in Figure 5, in which the well-ordered mesoporous structure typical for SBA-15 can be distinguished, indicating that the mesoporous structure is preserved. This observation aligns with the N₂ physisorption results obtained for the calcined samples. In all cases, dark zones can be distinguished on the SBA-15 channels corresponding to Co⁰ and CeO₂ particles, as corroborated by EDX. Furthermore, in the micrograph of Co/SBA-15 (Figure 5A), particles on the external surface of the SBA-15 are clearly distinguishable, suggesting some Co⁰ particle clustering, as reported before [52]. Moreover, not all of the SBA-15 surface or channels were filled with cobalt particles. On the other hand, in CeO₂-doped catalysts (Figure 5B–E), smaller crystallites can be differentiated within SBA-15 channels, corroborating the higher active phase dispersion expected from the mean crystallite sizes shown in Table 1.

From the TEM micrographs, particle size distribution histograms of the reduced catalysts were determined. For Co/SBA-15, the distribution was broader and centered at about 8 nm, indicating a higher particle size, in agreement with the agglomerate formation on the surface. With the CeO₂ incorporation, a decrease in the particle size was observed along with a narrower distribution, with values around 3 nm attained for the Co/30CeO₂-SBA-15 catalyst. This is consistent with the Co⁰ crystal size decrease determined by XRD (Table 1). Furthermore, even though Co⁰ and CeO₂ particles were not distinguishable in the TEM micrographs, both were included in the measurement distribution, given that the CeO₂ mean crystallite size ranges from 5.4 nm (Co/5CeO₂-SBA-15) to 7.3 nm (Co/30CeO₂-SBA-15), which overlaps with the distribution range.

3.2. Oxidative Steam Reforming of Acetic Acid

The effect of the CeO₂ content on the catalytic activity of Co/xCeO₂-SBA-15 catalysts in AAOSR was evaluated in the range of 400 to 700 °C. In this regard, Figure 6a,b display the results of catalytic activity concerning H₂ yield and acetic acid conversion as a function of the reaction temperature, respectively. As predicted, given the endothermic characteristics of the reforming reactions, an increase in the reaction temperature led to an improvement in the catalytic activity for all the catalysts [53]. Concerning the H₂ yield, low values were achieved at temperatures below 500 °C. However, increasing the reaction temperature from 500 °C to 550 °C significantly enhanced the H₂ yield, with values around 30% higher reached for the catalysts supported over modified CeO₂-SBA-15. Beyond this temperature, ceria-containing samples considerably increased the H₂ yield as compared to Co/SBA-15, reaching H₂ yields up to 70–75% in all cases, near thermodynamic equilibrium values. Therefore, the effect of Ce content was less remarkable with further temperature rises.

To analyze the efficacy of the incorporation of CeO₂ into the Co/SBA-15 catalysts, concerning acetic acid conversion, it is required to focus on temperatures below 550 °C. Over this range, the Co/5CeO₂-SBA-15 catalyst achieved approximately 5% greater acetic acid conversion than Co/SBA-15. Similarly, the acetic acid conversion increased by almost 20% for Ce contents over 10 wt.% in the catalyst formulation. Likewise, acetic acid conversion increased by over 40% when increasing the temperature from 500 to 550 °C, reaching almost complete acetic acid conversions for catalysts Co/10CeO₂-SBA-15 and Co/20CeO₂-SBA-15. At temperatures above 600 °C, almost complete conversion was reached for all catalysts. These findings align with those mentioned by Bie et al. [26], who stated that cerium’s incorporation onto Co/SBA-15 catalysts for CO oxidation activated the oxygen in the surface lattice of SBA-15 (Ce^3+/4+-O-Si) and generated oxygen vacancies. In this way, the behavior of CeO₂-based materials has been related to ceria redox properties, which can release, capture, and store oxygen according to the following redox reaction:

$${\mathrm{CeO}}_2\leftrightarrow {\mathrm{CeO}}_{2-\mathrm{x}}{+\mathrm{O}}_{\mathrm{x}}$$

(15)

where O_x is the lattice oxygen at the CeO₂ surface [51]. This fact can promote the activity performance toward reactions such as the WGS (Equation (3)), which may increase the hydrogen yield [34]. Moreover, this lattice oxygen could induce a decrease in carbon deposition, promoting coke gasification [16,51]. During the AAOSR tests, the Co/10CeO₂-SBA-15 and Co/20CeO₂-SBA-15 samples showed similar H₂ yield and acetic acid conversion values. In contrast, the Co/30CeO₂-SBA-15 catalyst showed a slightly lower catalytic performance than the latter ones. The catalytic activity observed in the Co/30CeO₂-SBA-15 catalyst can be attributed to its textural properties, and lower oxygen vacancy. The Co/30CeO₂-SBA-15 catalyst exhibits the smallest BET surface area (240 m²/g) and pore volume (0.50 cm³/g), thus indicating a partial blockage of the mesoporous structure, which may hinder the diffusion of reactants and products and limit the accessibility to active sites in comparison with Co/10CeO₂-SBA-15 and Co/20CeO₂-SBA-15. Consequently, Co/10CeO₂-SBA-15 and Co/20CeO₂-SBA-15 catalysts demonstrate potential as suitable catalysts for hydrogen production through AAOSR.

Product distribution includes not only hydrogen but also several by-products detected in the output stream. In this context, the relations between (a) acetone, (b) methane, (c) carbon dioxide, and (d) carbon monoxide selectivities and the reaction temperature are graphically displayed in Figure 7. Acetone (Figure 7a) is an intermediate product that mainly was obtained at temperatures between 400 °C and 550 °C. Acetone would have been formed by adding acetyl (CH₃CO*) and methyl (CH₃*) groups that come from the acetic acid decomposition on the Co surface, according to Li et al. [54]. Acetone selectivity was significantly enhanced after the addition of CeO₂ to the Co/SBA-15 catalyst, reaching values between 65% and 70% at 400 °C with Co/10CeO₂-SBA-15, Co/20CeO₂-SBA-15, and Co/30CeO₂-SBA-15 catalysts. These results align with the findings reported by Davidson et al. [55] and Basagiannis et al. [56], who experimented with CeO₂-supported catalysts in the steam reforming (Equation (1)) and ketonization of acetic acid (Equation (5)); they concluded that CeO₂ promotes the ketonization of acetic acid. The mechanism of acetic acid ketonization proposed by Bossola et al. [57] suggests that acetone formation takes place via the ketene intermediate, proceeding from acetic acid decomposition to the acyl group, which occurs on the CeO₂ support surface, predominating at low reaction temperatures (<250–350 °C). This would suggest that the acetic acid ketonization pathway has been thermodynamically unfavored, given that it is an exothermic reaction (ΔH_{25 °C} =16.7 KJ/mol). Moreover, by increasing the reaction temperature, the acetone selectivity decreases, dropping to near 0% above 550 °C alongside increasing CO₂ (Figure 7c) and CO (Figure 7d) selectivity, which could imply that acetone steam reforming is undergoing.

Accordingly, above 550 °C, a slight enhancement in CO₂ selectivity alongside a CO selectivity decrease was observed on Ce-doped catalysts. Conversely, in the Co/SBA-15 catalyst, the CO selectivity increased alongside a slight CO₂ selectivity decrease. One of the possible reaction pathways for the Co/SBA-15 catalyst is the reverse water–gas shift (RWGS), which was thermodynamically favored as the temperature increased. By contrast, Ce-doped catalysts would be kinetically favoring the gas–water shift reaction (WGS) (Equation (3)), as stated above. Specifically, CeO₂ incorporation enhances the oxygen vacancy density on the catalyst surface [51]; thus, these vacancies promote the WGS reaction efficiently, oxidizing CO into CO₂. In addition, the improved cobalt dispersion leads to a greater number of active sites, evenly distributed, which further facilitates the CO oxidation.

Finally, the methane selectivity (Figure 7b) remained at very low values (<5%) from 400 to 700 °C, suggesting that the thermal decomposition of acetic acid to methane (Equation (4)) is not a kinetically or thermodynamically relevant pathway. Sun et al. [58] reported that one possible pathway to methane generation is the hydrogenation of CH₃* species, proceeding from the breaking of C-C bonds, which could be kinetically disfavored when using cobalt-based catalysts as well as catalysts supported on metal oxides such as CeO₂. There has been speculation that this effect could be caused by the ease of cleavage of C-H bonds, thus limiting the formation of CH₃* species, or also by the oxidation of CH₃* species to CO_x due to the oxygen mobility.

Given that the Co/xCeO₂-SBA-15 catalyst with 10–20 wt.% of CeO₂ exhibited the best catalytic performance, its stability was evaluated through a long-term test at 550 °C.

3.3. Stability Test on Co/10CeO₂-SBA-15 and Co/20CeO₂-SBA-15 Catalysts

Avoiding catalyst deactivation caused by harsh operating conditions has been a primary issue in the steam reforming process. Hence, the development of a stable catalyst plays a main role in hydrogen formation via AAOSR. The catalytic activity of Co/10CeO₂-SBA-15 and Co/20CeO₂-SBA-15 catalysts, with respect to acetic acid conversion and H₂ yield, has been investigated at 550 °C and under the same reaction conditions for long time-on-stream periods (50 h). Figure 8 displays the results that were achieved. The initial H₂ yield for the Co/10CeO₂-SBA-15 catalyst was 66.8%, and it was slightly lower (62%) for the Co/20CeO₂-SBA-15 sample. These two catalysts showed a slightly decreasing yield throughout the reaction time (50 h), reaching approximately 55% at 50 h. Similarly, the acetic acid conversion curves started at around 98% for both catalysts and ended at approximately 85%. H₂ yields close to the thermodynamic equilibrium value (67.4%) and high conversion values were achieved. These results highlight their capacity to maintain suitable catalytic performance over a prolonged period, constituting a fundamental requirement for continuous hydrogen production.

To compare the findings of this study with the previously reported studies of catalytic systems for acetic acid oxidative steam reforming, the stability and catalytic performance were compared to recent data reported in the literature (

Table 3. Catalytic performance of several catalysts in acetic acid oxidative steam reforming.

Catalyst	T (°C)	S/C Ratio	O2/AcOH Ratio	TOS (h)	GHSV (mL/gcat. h)	XAcOH (%)	YH2 (%) a	Ref.
Ni/CeO2-m	500	2	0.075	5	25,000	97	54	[38]
Ni/La2O3-CeO2-m	500	2	0.150	5	25,000	97	54	[25]
10Ni/Sm2O3-MnO	700	4	0.280	10	51,000	100	62.5	[59]
Ni0.08Mg0.10Ti0.37O0.92±δ	700	4	0.280	10	50,700	100	68	[60]
15 wt% NiO-75 wt% TiO2-10 wt% YO1.5	700	4	0.280	10	51,000	100	65.5	[61]
17%Co/CeO2 R	600	4	0.280	10	37,260	100	62.5	[62]
Ni0.43Mg2.56CrO4.5±δ	700	4	0.280	10	25,550	100	62.5	[63]
Co-Rh/CeO2	600	2	0.075	25	25,000	100	70	[39]
Ni0.8La1.35Pr0.73O3.92±δ	700	4	0.280	50	50,600	100	62.5	[64]
Ni0.80Ce1.76Y0.35O4.85±δ	700	4	0.280	50	50,600	100	65	[65]
Ni-2Sm-Al-O	700	4	0.280	30	30,000	100	65	[37]
Co/10CeO2-SBA-15	550	2	0.075	50	25,000	85	62	This Work
Co/20CeO2-SBA-15	550	2	0.075	50	25,000	85	55

^a Standardized H₂ yield data by applying Equation (13).

). In this way, a reliable reference framework for analyzing the feasibility of the proposed catalysts was made available.

Table 3 shows that the Co/10CeO₂-SBA-15 and Co/20CeO₂-SBA-15 catalysts achieved notable H₂ yields, in spite of being tested under thermodynamically and kinetically less favorable conditions (550 °C). In the literature, most of the research on acetic acid oxidative steam reforming has focused on Ni-based catalysts, frequently operating in autothermal conditions at 700 °C [37,59,60,61,63,64,65], reaching H₂ yields of 62-65%. Although a straightforward comparison becomes difficult due to differences in experimental conditions, the use of a standardized H₂ yield expression allows a more objective assessment. A competitive performance is demonstrated by Co/10CeO₂-SBA-15 and Co/20CeO₂-SBA-15 catalysts. Furthermore, a comparison with our studies performed under similar conditions [25,38,39] showed that both catalysts exhibited excellent long-term stability, maintaining their activity over 50 h time-on-stream.

3.4. Characterization of Used Catalysts After Stability Tests

Coke formation is widely recognized as a main deactivation factor in reforming catalysts, along with sintering and/or oxidation [66]. Thus, the used catalysts, after 50 h of time-on-stream, were characterized through XRD, TEM, Raman spectroscopy, and TGA techniques. Figure 9 displays the XRD patterns of the two spent catalysts. The presence of a new diffraction feature at 2θ = 26.31° is noticeable, as compared with the calcined and reduced catalysts, attributed to graphitic carbon (JCPDS 01-074-2329), in both catalysts. The Co/20CeO₂-SBA-15 catalyst displayed a more intense diffraction peak, suggesting that a higher coke deposition was taking place during the reaction. Furthermore, despite the oxidizing atmosphere during the oxidative steam reforming reaction, no Co₃O₄ peaks were observed in either sample, indicating that the active Co⁰ species did not oxidize after 50 h of reaction. Similar to the calcined and reduced catalysts, the mean crystallite sizes of metallic Co species were calculated using the Scherrer equation, with mean crystallite sizes of 5.0 for Co/10CeO₂-SBA-15 and 4.2 nm for Co/20CeO₂-SBA-15 obtained. Taking into account the crystallite sizes of the reduced catalysts (Table 1), a small increase in the mean crystallite size can be inferred, suggesting a moderate sintering effect; according to Ochoa et al. [66], the sintering route is significantly affected by the size of the crystal.

The mass loss profiles together with the thermogravimetric derivative curves of each spent catalyst are displayed in Figure 10. In both samples, the mass percentage started decreasing when the temperature reached around 450 °C, with most of the weight loss in the temperature range 450–600 °C. The coke deposition rate was calculated as 7.2 mg_coke·g_cat⁻¹·h⁻¹ for the Co/10CeO₂-SBA-15 catalyst and 12.0 mg_coke·g_cat⁻¹·h⁻¹ for Co/20CeO₂-SBA-15. The DTG profiles display two peaks, a lower-intensity peak centered at 460–480 °C, and a high-amplitude and intensity peak at 550–570 °C. According to the literature, these peaks correspond to the oxidation of deposited carbon nanofilaments with differing ordering degrees. At low temperatures (460–480 °C), filamentous carbon species oxidation took place, which was related to the coke deposition on the metal surface. At higher temperatures (550–570 °C), carbon species were described as aged filamentous carbon, which would be deposited on the catalyst support [51,67]. Coke deposited on the metallic sites would block the active sites directly, affecting the catalytic activity. In contrast, coke deposited on the support would mainly affect the reactants’ diffusion. On both catalysts, the coke deposited on the support (higher temperature peak) was predominant. A higher intensity for Co/20CeO₂-SBA-15 was observed, indicating a high carbon total accumulation.

The spent samples were also characterized by TEM, as shown in Figure 11. The SBA-15 porous mesostructure was preserved after carrying out the catalytic activity tests. Well-dispersed dark rounded particles can be appreciated, which correspond to Co⁰ and CeO₂. Concerning coke deposition, nanofilamentous coke can be appreciated for both samples, in agreement with DTG results. According to the literature [66], this coke type is generally ascribed to transition metals, such as Co, Ni, or Fe.

The particle size distribution histograms were obtained from TEM micrographs (Figure 11). The histograms show a shift in the particle size distribution to higher particle sizes in comparison with the reduced catalysts. Particularly, the Co/20CeO₂-SBA-15 catalyst presents a higher frequency of particles above 16 nm in size, deviating with respect to the Gaussian distribution profile. These higher particle sizes are related to sintering phenomena caused by the coke filament formation. These filaments would cause the detachment of the Co⁰ crystals from the mesoporous structure, favoring their agglomeration [21,66].

Finally, Figure 12 displays the Raman spectra of the spent catalysts. These spectra exhibit two predominant bands: the D band (~1340 cm⁻¹) and the G band (~1590 cm⁻¹), both attributable to different carbon structures [68,69]. The ratio of these two bands, represented as the I_D/I_G ratio, has been commonly applied as an indicator of the degree of structural order in the deposited carbon species [68]. In this research, TEM and TGA/DTG analysis demonstrated that the coke was composed mainly of carbonaceous nanostructures, as filaments with different degrees of ordering. Such structures are associated with a higher intensity in the D band [68]; therefore, an increase in the I_D/I_G ratio could be associated with a greater proportion of them. Accordingly, the Co/20CeO₂-SBA-15 catalyst exhibited an I_D/I_G ratio of 1.4, higher than the 0.8 of Co/10CeO₂-SBA-15, suggesting a greater amount of filamentous carbon with a less-ordered structure and, possibly, more tendency for coke accumulation.

4. Conclusions

This study highlights the significant potential of CeO₂-doped Co/SBA-15 catalysts for AAOSR, as bio-oil aqueous phase representative compounds. The characterization findings revealed that Ce-doped catalysts preserved the well-ordered mesostructure of the SBA-15 support while increasing the catalyst’s reducibility. Additionally, the incorporation of CeO₂ improved cobalt dispersion, as evidenced by smaller crystallite sizes, thus enhancing the catalytic performance and reducing the coke formation rate. Catalysts with 10–20 wt.% CeO₂ exhibited the best catalytic performance, reaching H₂ yields close to thermodynamic equilibrium and nearly complete acetic acid conversion at 550 °C. Furthermore, these catalysts maintained the H₂ yield above 55% and high acetic acid conversion with a low coke formation rate. These findings underscore the importance of using CeO₂ as a promoter in the catalyst formulation, highlighting that Co/10CeO₂-SBA-15 and Co/20CeO₂-SBA-15 catalysts are promising for reforming purposes.